Relationship-based search and recommendations via authenticated negatives

Abstract

One embodiment of the present invention sets forth techniques for generating recommendation sets for a first client device. A recommendation system receives, from the first client device, a first selection of a first recommended item included in a plurality of recommended items. The recommendation system identifies a second recommended item included in the plurality of recommended items that has not been selected. The recommendation system retrieves an authenticated negative item from a plurality of authenticated negative items. The recommendation system stores one or more entries in a log file comprising a plurality of entries, based on at least one of the first recommended item, the second recommended item, and the authenticated negative item. One advantage of the disclosed techniques is that the use of authenticated negative examples, also referred to herein as authenticated negative items, provides a more relevant set of recommendations for the user.

Description

BACKGROUND OF THE INVENTIONField of the Invention[0001]The present invention generally relates to computer software and, more particularly, to improved relationship-based search and recommendations via authenticated negatives.Description of the Related Art[0002]Numerous websites offer items, such as various goods and services, available for online selection or purchase by a user of the website. Such websites may present recommendations for consideration by the user. The recommendations may be based on a reinforcement learning model, whereby a processor executing a software agent makes recommendations based on past history, such as items previously selected or purchased by the user. In typical implementations, the software agent collects data related to which goods and services previously selected by the user and which goods and services previously recommended to the user were not selected by the user. The collected data from these past recommendations and selections informs decision...

AI Technical Summary

Benefits of technology

The described techniques make use of authenticated negative examples, which helps to provide more relevant recommendations for users. This results in a more meaningful experience for users.

Problems solved by technology

One potential drawback with the above approach is that non-selection of an item does not necessarily indicate that the user completely disfavors the non-selected item.

However, this possible solution results in the user receiving recommendations that are known to be unrelated to the user's previous history and are likely to be undesirable to the user.

As a result, the user's online experience with the website is diminished, thereby increasing the likelihood that the user visits alternate websites that present more desirable recommendations for selection and purchase of goods and services.

Images